G G reat science and good theories can be elegant and wonderful, but they are of little practical value until they can translate into real products or processes. This is especially true in the biopharmaceutical industry. Now that the human genome has been sequenced, many people outside the industry assume that knowing of a gene’s existence means that the associated protein has been found and the therapy should be available already. Scientists and those in the biotechnology industry realize that the truth is not so simple. A key detail is posttranslational modification of proteins within eukaryotic cells — perhaps the most important of those being the process of glycosylation, the addition of carbohydrate (sugar) side chains that dramatically affect the way each protein molecule is routed and processed within the body. For protein drugs, a glycosylation pattern can spell the difference between how quickly and well the drug works — or whether it works at all — and can have enormous effects on manufacturing yields and cost of production. Traditionally, therapeutic glycoproteins have been produced through the somewhat costly route of inserting a gene for the target protein into a mammalian expression system such as Chinese hamster ovary (CHO) or baby hamster kidney (BHK) cells. Recombinant bacteria produce proteins less expensively than mammalian cells, but they do not glycosylate at all; yeast and fungi add complex mannose structures, whereas plant cells add sugars such as xylose in chains where they never occur in mammalian proteins. Even CHO cell glycosylation is sometimes incomplete, resulting in lower efficacies, increased risk of immunogenicity, and reduced product uniformity. Gaining control of the glycosylation process not only provides a route to more Building a GMP Bacterial and Fungal Fermentation Facility A Case Study in Designing for Evolving Technologies Kevin Pelin, Karen Phillips, and Victor Sarantschin PRODUCT FOCUS: PROTEIN THERAPEUTICS PROCESS FOCUS: FACILITY DESIGN AND CONSTRUCTION WHO SHOULD READ: R&D, PROJECT MANAGERS, PROCESS DEVELOPMENT AND MANUFACTURING, FACILITY DESIGN AND ENGINEERING, SENIOR MANAGEMENT KEYWORDS: CONTRACTORS, SCALE-UP , FACILITIES, FERMENTATION, CONTROLLED ENVIRONMENTS LEVEL: INTERMEDIATE Setting up the 1,500-L bacterial fermentor NEOSE TECHNOLOGIES, INC. (WWW.NEOSE.COM) 56 BioProcess International JUNE 2003
Transcript
GG reat science and goodtheories can be elegantand wonderful, butthey are of littlepractical value until
they can translate into real productsor processes. This is especially truein the biopharmaceutical industry.Now that the human genome hasbeen sequenced, many peopleoutside the industry assume thatknowing of a gene’s existence meansthat the associated protein has beenfound and the therapy should beavailable already. Scientists andthose in the biotechnology industryrealize that the truth is not sosimple. A key detail isposttranslational modification ofproteins within eukaryotic cells —perhaps the most important of thosebeing the process of glycosylation,
the addition of carbohydrate (sugar)side chains that dramatically affectthe way each protein molecule isrouted and processed within thebody. For protein drugs, aglycosylation pattern can spell thedifference between how quickly andwell the drug works — or whetherit works at all — and can haveenormous effects on manufacturingyields and cost of production.
Traditionally, therapeuticglycoproteins have been producedthrough the somewhat costly routeof inserting a gene for the targetprotein into a mammalianexpression system such as Chinese
hamster ovary (CHO) or babyhamster kidney (BHK) cells.Recombinant bacteria produceproteins less expensively thanmammalian cells, but they do notglycosylate at all; yeast and fungiadd complex mannose structures,whereas plant cells add sugars suchas xylose in chains where they neveroccur in mammalian proteins. EvenCHO cell glycosylation is sometimesincomplete, resulting in lowerefficacies, increased risk ofimmunogenicity, and reducedproduct uniformity. Gaining controlof the glycosylation process not onlyprovides a route to more
Building a GMP Bacterial andFungal Fermentation FacilityA Case Study in Designing for Evolving Technologies
Kevin Pelin, Karen Phillips, and Victor Sarantschin
Setting up the 1,500-L bacterial fermentor NEOSE TECHNOLOGIES, INC. (WWW.NEOSE.COM)
56 BioProcess International JUNE 2003
economical production ofglycoproteins, but also opens thepossibility of modifyingglycosylation patterns to yield drugswith improved and more reliabletherapeutic or pharmacokineticprofiles.
Neose Technologies, Inc.(Horsham, PA, www.neose.com),was founded in 1990 based on aproprietary enzymatic synthesistechnology that makes possible theaccurate, large-scale commercialproduction of complexcarbohydrates. In 2000, thecompany launched theGlycoAdvance in vitro proteinglycosylation technology, whichincludes a toolkit of recombinantenzymes (glycosyltransferases) toprecisely, independently, andconsistently add the appropriatesugars to the correct locations onproteins after cellular expression.Sugars can also be conjugated topolyethylene glycol (PEG) usingGlycoPEGylation technology, whichallows for site-specific attachment ofPEG to glycans remote from theactive site of a protein. Thatprovides a way to maintainbiological activity while extendingmolecular half-life and reducingpotential immunogenicity for certaintherapeutics.
BUILDING TO EXPAND
With our set of proprietarytechnologies serving core needs ofthe rapidly growingbiopharmaceutical industry, we atNeose have faced expansionchallenges that are unique in theirparticulars but in some respectssimilar to those experienced bymany others in the field. Ouroriginal pilot plant, located adjacentto company headquarters in acorporate park outside ofPhiladelphia, could remodelglycoproteins in the gram range. Toenable production of industrialquantities of the glycosyltransferasesand sugar nucleotides required forour type of protein glycosylation, weneeded a considerably largerproduction capacity. In February2001, we began to design newfacilities to accommodate a
coordinated flow (and sometimessegregation) of the diverse processesemployed to create protein drugs ina GMP environment.
A Unique Facility: One of the mostunusual challenges our project teamfaced was a need for both bacterialand fungal fermentation to produceglycosyltransferases. Bacterial andeven mammalian cell expressionsystems are common in the biotechpharmaceutical industry, but fungalsystems (which can give remarkablyhigh yields) have traditionally beenlimited to commodities likedetergent enzymes. Consequently,there is little prior experience inGMP fungal production ofpharmaceutical grade enzymes. Ourteam was challenged to make themost of shared resources in a45,000-ft2 facility that includesresearch and developmentlaboratories and corporate officesand must rigorously control againstany possible cross-contamination ofcultures through water, air handling,or other systems — or throughmechanical spread and human error.Fungal spores can easily destroy abacterial culture, and bacteria caninfect and ruin a fungal productionrun.
Once the Neose managementteam agreed that its existing GMPpilot manufacturing facilities weretoo small to be remodeledadequately to provide both theneeded capacity and the required
separation of fermentation methods,construction of a new add-onbuilding was made an aggressivelyfast track project. Duringconstruction of the original pilotplant in 1997, we had experienced anumber of troubling disconnectsbetween the contractors whodesigned and installed theequipment and those who validatedit. The decision was therefore madeto look at well-integratedconstruction management firms thatcould handle everything fromdesign through constructionmanagement and validation. Webegan with six firms, whittled themdown to two, and made the finaldecision based on personalinteractions — that is, on withwhom we felt we could spend twoyears of hard work and intimatecontact under what would doubtlessat times be stressful circumstances.
Scope and Complexity: The job,known as the GlycoAdvance Facility(GAF) project, was led by aninternal project team of sixemployees: a project manager andrepresentatives from qualityassurance, corporate, engineering,and validation, with others broughtin at various stages along the way.The overall time commitment,particularly from the manufacturingand engineering groups, wastremendous for a company ofNeose’s size, making speed andefficiency all the more important.
An early conceptual plan soughtto meet the multiple needs withoutexpanding the building. Under thatconstraint, capacity was limited to150-L fermentation volumes. Ourrevised plan was to remodel aboutone-third of the existing pilot plantand to add 5,000 ft2 of juxtaposednew space at a total estimated costof $17 million, creating a total of20,000 ft2 of processing space and3,500 ft2 of utility areas. Thisincreased the scope of the projectdramatically but also presentedunanticipated opportunities. Freedof the constraint of working withinthe preexisting pilot plant’s shell, weelected to install 1,500-L workingvolume fermentors in addition to
JUNE 2003 BioProcess International 57
The decision wasmade to look atwell-integratedconstructionmanagement firmsthat could handleEEVVEERRYYTTHHIINNGGfrom designthroughconstruction andvalidation.
����
58 BioProcess International JUNE 2003
the 15-L and 150-L fermentorsoriginally planned. The result wasthat, for a 35% increase in design,construction, and equipment costs,we were able to multiply productioncapacity tenfold.
One of the great challenges inthis project was that of designingthe facility for multiple projects withvarying degrees of complexity. Someprocesses for which the plant wasbeing built were well understoodand required scale-up, but otherswere still at the benchtop researchstage. The company’s business planwas in a dynamic state, with both
sugar nucleotide and fermentationprocesses comprising new productopportunities. Thus, settingpriorities and scheduling theconstruction of the variousmanufacturing suites was a challengebecause of the lead times involved inordering key large equipment.Ultimately, we chose to design thefacility for maximum flexibility toaccommodate any eventuality.
Automation: The decision wasmade early to create a highlyautomated facility. The fermentors,harvest sequence, and even clean-in-place (CIP) and steam-in-place(SIP) cycles would be controlled bya supervisory control and dataacquisition (SCADA) system on itsown dedicated server. Thisnecessitated working with vendorswho could provide packages thatmet the evolving FDA policies onelectronic record keeping and wouldcreate traceable, verifiable electroniclogs of every system activity. Thiscomplex control system was acritical path item that dictated theschedule of the project and otheractivities.
THE FINAL DESIGN
The sugar nucleotide synthesis areawas designed as a class 100,000
environment, including multiplereaction suites with vessels up to1,200 L in size for enzymaticsynthesis of carbohydrates anddownstream processing suites withvessels of up to 4,000 L in size andequipment for filtration andchromatographic purification.
The bacterial and fungalfermentation areas are virtuallyidentical facilities, each with adedicated inoculum laboratory; fully automated 15-L, 150-L, and1,500-L working volumefermentors; a harvest suite withhomogenizers and centrifuges all ina class 100,000 environment; and adownstream class 10,000environment purification suite.
Harvesting needed to be wellsegregated from the rest of theproduction areas to control anyaerosols created during thehomogenization and separationprocesses. Final purification andpackaging are carried out in an evencleaner area because theglycosyltransferase enzymesproduced in this facility will come indirect contact with activepharmaceutical products and musttherefore meet rigorous regulatorystandards.
Utilities: Air handling systems areseparate for the three mainproduction areas, but otherwise weattempted to share as many utilitiesas possible. The original pilot plantemployed mobile CIP systems, butwith 1,500-L fermentors it seemedmore reasonable to installcentralized CIP built into the utilityareas and designed to serve everyfermentor, harvest tank, andcentrifuge. Other shared utilitysystems include USP water, cleansteam, clean air, hot and coldprocess water, and plant steam.
Not surprising to those who havegone through such a multifaceteddesign and construction project, lessthan half of the cost of our newfacility was represented by the actualbuilding and major equipment. Mostof the cost — and most of thecomplexity — lay in the walls and theplenum space and under the floors.The various air, water, CIP, electrical,HVAC, process piping, and data lines
1,500-L fermentor (center left) and control screen (left) and 2,000-L harvest tank(right) NEOSE TECHNOLOGIES, INC. (WWW.NEOSE.COM)
For a 35% increasein design,construction, andequipment costs,we were able tomultiply ourproductioncapacityTTEENNFFOOLLDD..
����
had to be organized and layeredaccording to permanence and serviceneeds. Access panels for circuitbreakers, shutoff valves, filterchangers, and the like were situatedin corridors or outside the pilot plantenvelope whenever possible to permitservicing without breaking theenvironmental integrity of criticalareas.
Validation and Qualification:Meanwhile, the development ofvalidation protocols began early inthe overall process, with theintention of making the vendors’factory acceptance testing flow asseamlessly as possible into thevalidation program itself. However,no amount of expertise or off-linetesting can ever accurately predicthow a system will work whenchallenged with the first large-scalerun of a particular product.
Once remodeling andconstruction were complete, theequipment was installed andcommissioned. There wereequipment turn-over packages foreach item, and 112 writtenprotocols made up the validationpackage.
One significant delay in theconstruction and outfitting(suggestion that in spite of the mostcareful plans, something can alwaysgo awry) was caused by discussionswith the local authorities over theneed for a storm-water managementsystem to reduce runoff from oursite. A few months after plant designbegan, the township was hit by aheavy summer storm that causedsignificant flooding of businessesand residences in the immediateneighborhood. The resulting storm-water runoff issues set ourconstruction project back by aboutfour months. This was an untimelyreminder of the importance of localzoning and approval issues. Tominimize the unavoidable time loss,remodeling of our existing pilotplant continued with some pipingstubbed into the walls awaitingconstruction of the building itwould eventually serve.
PLANNING IS KEY
Overall, design of our remodeled
and expanded facility began in lateFebruary 2001, demolition tookplace during that summer, andconstruction was essentiallycomplete by the fall of 2002.Installation qualification (IQ) beganas soon as systems werecommissioned and continued duringthe final phase of equipmentinstallation in the new structure.The design, implementation, andconstruction of our GMP reagentfacility was a challenging experiencebecause we were planning a facilitythat would house both bacterial andfungal expression yet allow for theflexibility of both internal andexternal requirements.
Many of the lessons we learnedseem intuitive in retrospect,although the answers were notalways obvious at the time.Certainly, selection of engineering,construction, and validationproviders is a vitally importantprocess. Equally important isunderstanding the scope ofvalidation and clarifyingresponsibilities and deliverables asearly as possible. Rigorous timelinesmust be created, and it is necessaryto have frequent and opencommunication between yourcompany and its engineering,construction, and validation partnersto keep to those timelines or planaround any unforeseen butinevitable contingencies. Thesuccess of a facility constructionproject depends upon it. ��
Kevin Pelin is vice president ofmanufacturing operations, KarenPhillips is a process engineer III andserved as validation manager on theproject, and corresponding authorVictor Sarantschin is director oftechnical marketing and businessdevelopment for Neose Technologies,Inc., 102 Witmer Road, Horsham, PA19044, 215-315-9000, fax 215-315-9100, [email protected].,www.neose.com. Neose is abiopharmaceutical company focused onimproving protein therapeutics throughits proprietary (GlycoAdvance andGlycoPEGylation) technologies.
As a Gas SpargerUsing the Pharmenta SIPTube® valve asa gas sparger you can assure full aerationcapacity during critical periods by reclean-ing and resterilizing sparger element poresin place as needed. Changeable sparger ele-ment lets you optimize for specific needs.
One Valve, Multiple UsesThe Pharmenta SIPTube® valve canreplace five traditional devices, but with ahuge improvement over each: the abilityto reclean and resterilize anytime – evenduring production and even if the processis heat sensitive.The result:eliminate dead-space, clogging,growback andother contami-nation and capacity nightmares.
Patented InnovationTip-Seal™ technology eliminates deadspaceby positioning the valve seal on the shafttip at the process interface instead ofoutside the vessel. Clean-Thru/Steam-Thru™ designmeans the entire valve passage forms a deeploop – from inlet to tip-seal to outlet – thatcan be recleaned and resterilized anytime. Optional insulation protects heat sensi-tive processes during steam sterilization. Weld-in or retrofit to existing ports assmall as 1 inch sensor ports. Designs are available to work in anyorientation. To learn more call 800-765-7838 orvisit www.pharmenta.com.
